Method and system for dynamically updating high-precision map of unmanned vehicle

Abstract

The invention discloses a high-precision map dynamic updating method and system for an unmanned vehicle, and relates to the technical field of map updating. The method comprises the following steps: analyzing a map updating time sequence of a current driving route according to a current planned route of a target vehicle in combination with a current driving speed; setting collaborative connection conditions; the cooperative updating vehicles are connected, spatial position alignment and time sequence arrangement are carried out according to cooperative front time zones of the cooperative updating vehicles, and cooperative weights of the cooperative updating vehicles are configured; and carrying out feature aggregation on the obtained dynamic map of the collaboratively updated vehicle according to the collaboration weight, the spatial position alignment and the time sequence arrangement relationship to obtain map updating data, and carrying out map dynamic updating by utilizing the map updating data. The technical problems of poor real-time performance and inaccurate data of the high-precision map updating of the unmanned vehicle in the prior art are solved, and the technical effects of improving the map updating precision and real-time performance and enhancing the driving safety and reliability of the unmanned vehicle are achieved.

Description

Technical Field

[0001] The present invention relates to the field of map updating technology, and in particular to a method and system for dynamically updating high-precision maps of unmanned vehicles. Background Art

[0002] With the rapid development of driverless technology, high-precision maps have become the key foundation for driverless vehicles to achieve core functions such as autonomous navigation, path planning and obstacle avoidance. High-precision maps not only provide detailed road geometry information, such as lane lines, traffic signs, intersection layouts, etc., but also contain rich semantic information, such as road type, speed limit, traffic signal status, etc., providing accurate environmental perception and decision support for driverless vehicles. However, the road environment in the real world is dynamically changing, including road construction, traffic sign replacement, temporary closures, etc. If these changes are not reflected in high-precision maps in a timely manner, they will pose a hidden danger to the driving safety and reliability of driverless vehicles. Therefore, how to achieve dynamic updates of high-precision maps and ensure the real-time and accuracy of map information has become an important issue that needs to be urgently solved in the current field of driverless technology. Summary of the invention

[0003] The present application provides a method and system for dynamically updating high-precision maps of unmanned vehicles, which solves the technical problems of poor real-time update performance and inaccurate data of high-precision maps of unmanned vehicles in the prior art.

[0004] In view of the above problems, the present application provides a method and system for dynamically updating high-precision maps of unmanned vehicles.

[0005] In a first aspect of the present application, a method for dynamically updating a high-precision map of an unmanned vehicle is provided, the method comprising: According to the current planned route of the target vehicle and the current driving speed, the map update timing of the current driving route is analyzed, and the map update timing is the map update time window of each road section on the planned route; according to the current planned route and its map update timing, the collaborative connection conditions are set, and the collaborative connection conditions include the collaborative vehicle line and the collaborative leading time zone, and the collaborative leading time zone is used to characterize the leading driving space-time area of the collaborative vehicle; based on the collaborative connection conditions, the collaborative update vehicles are connected, the spatial position alignment and time sequence arrangement are performed according to the collaborative leading time zone of the collaborative update vehicles, and the collaborative weights of the collaborative update vehicles are configured, and the collaborative weights correspond to the offset differences of the spatial position alignment and time sequence arrangement; according to the collaborative weights, the dynamic map acquired by the collaborative update vehicles is feature aggregated according to the spatial position alignment and time sequence arrangement relationship to obtain map update data, and the map update data is used to dynamically update the map.

[0006] In the second aspect of the present application, a high-precision map dynamic update system for driverless vehicles is provided. The system includes: An analysis module: Analyze the map update timing of the current driving route according to the current planned route of the target vehicle combined with the current driving speed. The map update timing is the map update time window for each section of the planned route; A condition setting module: Set collaborative connection conditions according to the current planned route and its map update timing. The collaborative connection conditions include collaborative vehicle routes and collaborative pre-time zones, and the collaborative pre-time zones are used to represent the pre-driving space-time regions of collaborative vehicles; A configuration module: Connect collaborative update vehicles based on the collaborative connection conditions, align the spatial positions and arrange the time sequences according to the collaborative pre-time zones of the collaborative update vehicles, and configure the collaborative weights of the collaborative update vehicles. The collaborative weights correspond to the offset differences in spatial position alignment and time sequence arrangement; A map update module: Aggregate the features of the dynamic maps obtained by the collaborative update vehicles according to the collaborative weights according to the spatial position alignment and time sequence arrangement relationships to obtain map update data, and use the map update data for map dynamic update.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages: First, analyze the map update timing of the current driving route according to the current planned route of the target vehicle combined with the current driving speed. The map update timing is the map update time window for each section of the planned route. Then, set collaborative connection conditions according to the current planned route and its map update timing. The collaborative connection conditions include collaborative vehicle routes and collaborative pre-time zones, and the collaborative pre-time zones are used to represent the pre-driving space-time regions of collaborative vehicles. Then, connect collaborative update vehicles based on the collaborative connection conditions, align the spatial positions and arrange the time sequences according to the collaborative pre-time zones of the collaborative update vehicles, and configure the collaborative weights of the collaborative update vehicles. The collaborative weights correspond to the offset differences in spatial position alignment and time sequence arrangement. Finally, aggregate the features of the dynamic maps obtained by the collaborative update vehicles according to the collaborative weights according to the spatial position alignment and time sequence arrangement relationships to obtain map update data, and use the map update data for map dynamic update. This solves the technical problems of poor real-time performance and inaccurate data in the high-precision map update of existing driverless vehicles, and achieves the technical effects of improving the map update accuracy and real-time performance, and enhancing the driving safety and reliability of driverless vehicles. Description of the Drawings

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0009] Figure 1 Schematic flowchart of the high-precision map dynamic update method for the driverless vehicle provided by the embodiment of the present application; Figure 2 Schematic structural diagram of the high-precision map dynamic update system for the driverless vehicle provided by the embodiment of the present application.

[0010] Explanation of reference numerals: analysis module 11, condition setting module 12, configuration module 13, map update module 14. Detailed implementation manners

[0011] By providing the high-precision map dynamic update method and system for the driverless vehicle, the present application solves the technical problems of poor real-time performance and inaccurate data in the high-precision map update of the existing driverless vehicle.

[0012] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0013] It should be noted that the terms "include" and "have" are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server including a series of steps or units does not necessarily have to be limited to those clearly listed steps or units, but may include other steps or modules not clearly listed or inherent to these processes, methods, products or devices.

[0014] Embodiment 1, as Figure 1 shown, the present application provides a high-precision map dynamic update method for a driverless vehicle, wherein the method includes: Analyze the map update timing of the current driving route according to the current planned route of the target vehicle in combination with the current driving speed, and the map update timing is the map update time window for each section on the planned route.

[0015] The current planned route of the target vehicle refers to the road or path that the vehicle plans to drive during the current journey; the current driving speed refers to the current motion state of the vehicle, which affects the driving time of the vehicle and the frequency of map data update. The faster the speed, the shorter the time for the vehicle to pass through each section, and the map update time window may also need to be adjusted accordingly.

[0016] The system analyzes the map update timing sequence of each section in the route based on the current planned route and the current driving speed of the target vehicle. The map update timing sequence refers to the map update time window for each section, and this window determines when the map data of this section needs to be updated. These time windows can be dynamically calculated and adjusted according to factors such as the vehicle's driving speed, the current route plan, and road conditions. For example, if the target vehicle is driving at a high speed, the map update frequency may increase to ensure that the map data can timely reflect the changes in road conditions.

[0017] Furthermore, the analysis of the map update timing sequence of the current driving route based on the current planned route of the target vehicle combined with the current driving speed includes: Extract the section nodes of the current planned route and identify the length of each section; calculate the estimated passing time of each section node according to the current driving speed and the length of each section; arrange the section nodes in time sequence and at time interval distances according to the estimated passing time of each section node to obtain the map update timing sequence of the current driving route.

[0018] Specifically, extract multiple section nodes from the current planned route of the target vehicle. Each section node represents an important geographical location during the driving of the target vehicle, such as intersections, crossroads, etc. For each section node, identify and record the length of the section (i.e., the distance between two adjacent nodes); calculate the estimated passing time of each section based on the current driving speed of the target vehicle, where the estimated passing time = section length / current driving speed; arrange the sections in time sequence according to the estimated passing time of each section; based on the time sequence arrangement of the section nodes, obtain the map update timing sequence of the current driving route.

[0019] Furthermore, obtaining the map update timing sequence of the current driving route includes: Calculate the adjustment time threshold of each section node according to the length of the section and the current driving speed; set the window tolerance duration according to the change penalty coefficient of each section node in the current planned route, where the change penalty coefficient is determined according to the section length and the route adjustment difficulty; perform constraint adjustment on the adjustment time threshold according to the window tolerance duration to determine the map update time window of each section node and obtain the map update timing sequence of the current driving route.

[0020] Specifically, according to the current driving speed of the target vehicle and the length of each road segment, the adjustment time threshold for each road segment is calculated. The adjustment time threshold refers to the time interval within a certain spatio-temporal range that allows adjustments to map updates, and the adjustment time threshold is equal to the road segment length divided by the current driving speed. According to the change penalty coefficient of each road segment, the window tolerance duration is set. The change penalty coefficient is calculated based on the road segment length and the difficulty of route adjustment. For example, longer road segments and complex traffic conditions will result in larger change penalty coefficients. Therefore, the window tolerance duration will be appropriately extended to ensure sufficient flexibility in the update process. Among them, the window tolerance duration = change penalty coefficient × adjustment time threshold. After determining the window tolerance duration, the adjustment time threshold for each road segment will be constrained and adjusted based on this tolerance duration. According to the set window tolerance duration, the original adjustment time threshold will be appropriately extended or shortened to ensure that map updates can be carried out within the most appropriate time, avoiding the occurrence of situations where the time window is too small or too large. According to the adjusted time threshold and the window tolerance duration, the map update time window for each road segment node is determined, that is, the map update time sequence of the current driving route is obtained.

[0021] According to the current planned route and its map update time sequence, the collaborative connection conditions are set. The collaborative connection conditions include the collaborative vehicle route and the collaborative pre-time zone, and the collaborative pre-time zone is used to represent the pre-driving spatio-temporal area of the collaborative vehicle.

[0022] According to the current planned route of the target vehicle, the collaborative vehicle routes adjacent to or intersecting with the target vehicle's driving route are identified. The collaborative vehicle routes refer to those other vehicle driving routes that are connected or closely related to the target vehicle's driving route. Vehicles on these routes may drive within the same spatio-temporal range and may affect the map update of the target vehicle.

[0023] The collaborative pre-time zone is used to represent the pre-driving spatio-temporal area of the collaborative vehicle. Specifically, the collaborative pre-time zone refers to the area within a certain time or space range in front during the driving process of the collaborative vehicle. The collaborative pre-time zone includes the road segments or areas that the collaborative vehicle is about to pass through, and the map data of these areas needs to be docked with the map update time sequence of the target vehicle. The setting of the collaborative pre-time zone depends on multiple factors, including the driving speed, route, and driving time sequence of the collaborative vehicle. For example, a collaborative vehicle with a faster driving speed may have a shorter pre-time zone, while a vehicle with a slower speed may have a longer pre-time zone. Through the comprehensive analysis of these factors, the system can reasonably predict the influence range of the collaborative vehicle on the map update of the target vehicle.

[0024] The collaborative connection conditions are set by combining the collaborative vehicle routes and the collaborative pre - time zones to ensure that the map update data between the target vehicle and the collaborative vehicles can be synchronized within an appropriate spatio - temporal range. By analyzing the driving timings, positions, and pre - time zones of the target vehicle and the collaborative vehicles, the collaborative connection conditions are flexibly adjusted to ensure that the collaborative vehicles can provide necessary map data support in a timely manner, improving the accuracy and real - time performance of map updates.

[0025] Connect the collaborative update vehicles based on the described collaborative connection conditions, align the spatial positions and arrange the timings according to the collaborative pre - time zones of the collaborative update vehicles, and configure the collaborative weights of the collaborative update vehicles, where the collaborative weights correspond to the offset differences in spatial position alignment and timing arrangement.

[0026] Based on the collaborative connection conditions, the system connects the collaborative update vehicles with the target vehicle, and performs spatial position alignment and timing arrangement according to the collaborative pre - time zones of the collaborative update vehicles. Determine the collaborative update vehicles related to the target vehicle according to the collaborative connection conditions, and ensure that the data of these collaborative vehicles can effectively cooperate with the map update requirements of the target vehicle. In terms of spatial position alignment, the system aligns the driving trajectories of the collaborative vehicles with the map update areas of the target vehicle according to the current positions of the collaborative vehicles and the map update requirements of the target vehicle, including calculating the relative spatial positions between the collaborative vehicles and the target vehicle and making appropriate adjustments to ensure that the collaborative vehicles can provide map update data at the correct positions. In terms of timing arrangement, the system makes a reasonable arrangement according to the driving timings of the collaborative vehicles to ensure that the map update timings of different collaborative vehicles can accurately match the map update timings of the target vehicle. Through this timing arrangement, the system can ensure that the map data of different collaborative vehicles is integrated in the appropriate time order to improve the real - time performance and accuracy of map updates.

[0027] The system configures a collaborative weight for each collaborative update vehicle. The collaborative weight represents the importance of the collaborative vehicle in the map update process and is associated with the offset differences in its spatial position alignment and timing arrangement. Specifically, the magnitude of the weight depends on the spatial alignment accuracy and timing matching degree between the collaborative vehicle and the target vehicle. The smaller the offset differences in spatial position alignment and timing arrangement, the larger the collaborative weight, and vice versa.

[0028] Furthermore, connect the collaborative update vehicles based on the described collaborative connection conditions, align the spatial positions and arrange the timings according to the collaborative pre - time zones of the collaborative update vehicles, and configure the collaborative weights of the collaborative update vehicles, including: Obtain the pre - space position and pre - time relationship between the collaborative update vehicle and the target vehicle; align the coordinates of the target vehicle and the collaborative update vehicle according to the pre - space position to determine the position alignment amount; calculate the time deviation amount between the collaborative update vehicle and the target vehicle according to the pre - time relationship; calculate the space alignment weight and the time alignment weight based on the position alignment amount and the time deviation amount, and perform weight fusion according to a preset fusion coefficient to obtain the collaborative weight value of the collaborative update vehicle.

[0029] Specifically, obtain the pre - space position and pre - time relationship between the collaborative update vehicle and the target vehicle. The pre - space position refers to the geographical position of the collaborative vehicle in the front, and the pre - time relationship is the difference in the driving time sequence between the collaborative vehicle and the target vehicle. Based on the obtained pre - space position relationship, the system aligns the coordinates of the target vehicle and the collaborative update vehicle, that is, by calculating the spatial offset between the two, the spatial position of the collaborative vehicle is adjusted to align it with the target vehicle in space. The aligned position alignment amount reflects the degree of alignment in space between the two. The smaller the position alignment amount, the more precisely the two positions are aligned. On the basis of the spatial position alignment, calculate the time deviation amount according to the pre - time relationship. The time deviation amount refers to the difference between the time node of the collaborative vehicle and the time node of the target vehicle. By calculating this difference, the system can judge the time synchronization degree of the collaborative vehicle and the target vehicle. The smaller the time deviation amount, the more precisely the collaborative vehicle and the target vehicle are docked in time sequence. According to the calculated position alignment amount and time deviation amount, calculate the space alignment weight and the time alignment weight respectively. The space alignment weight is determined by the spatial alignment amount between the two. The smaller the spatial alignment amount, the greater the weight. Similarly, the time alignment weight is calculated according to the time deviation amount. The smaller the time deviation amount, the greater the weight. The space alignment weight and the time alignment weight are fused according to the preset fusion coefficient to obtain the collaborative weight value of each collaborative update vehicle.

[0030] Furthermore, obtain the collaborative weight value of the collaborative update vehicle, and its calculation expression is: , where is the collaborative weight value of collaborative update vehicle i, is the fusion coefficient of the space alignment weight, is the fusion coefficient of the time alignment weight, is the position alignment deviation amount, is the time deviation amount, is the space alignment attenuation factor, is the time alignment attenuation factor.

[0031] The fusion coefficient of the space alignment weight represents the contribution degree of the space alignment to the collaborative weight value; is the fusion coefficient of the time alignment weight, representing the contribution degree of time alignment to the collaborative weight; is the position alignment deviation amount, representing the deviation between the collaborative update vehicle and the target vehicle in terms of spatial position; is the time deviation amount, representing the deviation between the collaborative update vehicle and the target vehicle in terms of time; is the spatial alignment attenuation factor, which measures the attenuation effect of the spatial position difference between vehicles on the data contribution in the collaborative update task. The spatial alignment attenuation factor is used to adjust the influence of the position alignment deviation amount, determine the attenuation speed of the position deviation amount on the collaborative weight. The larger the attenuation factor, the more sensitive the influence of the position alignment deviation amount on the weight, and the greater the position deviation, the faster the weight decreases; is the time alignment attenuation factor, which measures how the time synchronization between vehicles (i.e., the difference in data acquisition time) affects their contributions. The time alignment attenuation factor is used to adjust the influence of the time deviation amount, determine the attenuation speed of the time deviation amount on the collaborative weight. The larger the attenuation factor, the more sensitive the influence of the time deviation amount on the weight, and the greater the time deviation, the faster the weight decreases. Through this calculation formula, the system can adjust the collaborative weight of the collaborative update vehicle according to the magnitudes of the spatial alignment deviation amount and the time deviation amount, and by means of the fusion coefficient and the attenuation factor, so as to achieve a reasonable evaluation of the contribution degrees of different collaborative vehicles.

[0032] Furthermore, obtaining the collaborative weight of the collaborative update vehicle further includes: Obtaining the record update database of the collaborative update vehicle, where the record update database is the historical acquisition data of the collaborative update vehicle, including different acquisition device tags; performing data credibility evaluation according to the record update database to obtain the data evaluation coefficients of each acquisition device; configuring the collaborative weights of the acquisition data sources of the collaborative update vehicle according to the data evaluation coefficients of each acquisition device, and adding the collaborative weights of the acquisition data sources to the collaborative weight of the collaborative update vehicle.

[0033] Preferably, obtain the record update database of each collaborative update vehicle. The record update database contains the historical data collected by the collaborative update vehicle through various acquisition devices in the past period of time. These data include multiple different types of acquisition device tags (such as GPS sensors, lidar, cameras, etc.). Each device may collect different types of geographical location information, environmental data or road features. After obtaining the record update database, evaluate the data credibility of each acquisition device according to the historical acquisition data, and calculate the data evaluation coefficient of each acquisition device according to the historical performance of each device (such as factors such as data accuracy, stability, acquisition frequency, etc.). These coefficients reflect the reliability of the data provided by each device. The larger the coefficient, the higher the credibility of the device data, and vice versa. According to the obtained data evaluation coefficients, the system configures a collaborative weight for each acquisition data source of the collaborative update vehicle. These collaborative weights are used to measure the weight that the acquisition data source should occupy in the entire map update process. Specifically, the acquisition data source with higher credibility will be given a higher collaborative weight, while the acquisition data source with lower credibility will be given a lower weight. Integrate the collaborative weight of each acquisition data source with the original collaborative weight of the corresponding collaborative update vehicle. By integrating the collaborative weights of the acquisition data sources, the system can comprehensively consider factors such as spatial alignment, time alignment, and device credibility to obtain a final collaborative weight.

[0034] According to the collaborative weight, perform feature aggregation on the acquired dynamic map of the collaborative update vehicle according to the spatial position alignment and the time sequence arrangement relationship to obtain map update data, and use the map update data to perform dynamic map update.

[0035] According to the collaborative weight, the system will perform feature aggregation on the acquired dynamic map of the collaborative update vehicle according to the spatial position alignment and the time sequence arrangement relationship. Feature aggregation is to integrate the map information of each collaborative vehicle to facilitate the creation of a more accurate and complete map update data. Dynamically update the map based on the map update data to ensure that the driverless vehicle can obtain the latest and most accurate map data.

[0036] Furthermore, according to the collaborative weight, perform feature aggregation on the acquired dynamic map of the collaborative update vehicle according to the spatial position alignment and the time sequence arrangement relationship to obtain map update data, including: Obtain the dynamic map feature data of the collaborative update vehicle, where the dynamic map feature data has a data source identifier; align the spatial positions and arrange the time series of the dynamic map feature data, and use the collaborative weights obtained from spatial alignment and time alignment to perform feature screening on the dynamic map feature data to obtain the map spatio-temporal features to be aggregated; calculate the contribution value of the feature data of each collaborative update vehicle through the attention mechanism according to the collaborative weights of the acquisition data source and the collaborative weights of the collaborative update vehicle; perform feature aggregation on the map spatio-temporal features to be aggregated according to the contribution value of the feature data of each collaborative update vehicle to obtain the map update data.

[0037] Specifically, obtain the dynamic map feature data of each collaborative update vehicle. These data contain map feature information obtained from different acquisition devices (such as GPS, lidar, cameras, etc.). The feature data of each data source is marked with a data source identifier to distinguish data from different sources. Then, the system performs spatial position alignment and time series arrangement on these dynamic map feature data. Spatial position alignment ensures that the data provided by different collaborative vehicles is aligned in a unified geographic coordinate system to avoid position deviation; time series arrangement ensures that all data is arranged in chronological order to maintain the freshness and timeliness of the data. After spatial alignment and time alignment are completed, the system performs feature screening on the data according to the collaborative weights. Through the setting of the weights, the system screens out the data from the collaborative vehicles with high weights, and these data will play a more important role in the aggregation process. Next, the system uses the attention mechanism to calculate the contribution value of the feature data of each collaborative update vehicle according to the collaborative weights of each collaborative update vehicle and the weights of each acquisition data source. Among them, the feature data of the vehicle refers to the relevant information provided by each collaborative update vehicle during the dynamic update of the map; through this mechanism, the data provided by the vehicle with a larger weight will have a greater impact on the final map update result. Finally, the system performs weighted aggregation on the map spatio-temporal features to be aggregated according to the calculated contribution value of the feature data. This process performs weighted fusion on the data of different collaborative vehicles to ensure that the data provided by each vehicle according to its contribution value occupies an appropriate proportion in the final map update. Finally, through these steps, the system obtains the merged, more accurate and comprehensive map update data, providing real-time and accurate map support for autonomous vehicles.

[0038] The attention mechanism refers to dynamically assigning different weights to different parts when processing data, so as to focus on the most important information of the task. In the feature aggregation stage, the contribution value of the feature data of each collaborative update vehicle is calculated through the attention mechanism, that is, the attention mechanism automatically adjusts the influence of the feature data of each vehicle according to the collaborative weights of each collaborative update vehicle and the collaborative weights of the acquisition data source.

[0039] By defining a scoring function , used to represent the correlation between each collaborative update vehicle and the data source, where is a scoring function, is the collaborative weight of the collaborative update vehicle, is the weight of the data source for data collection.

[0040] The scoring function is used to calculate the score between each collaborative update vehicle and the data source, and then the attention weight of each data source to the collaborative vehicle is obtained through a normalization process. The normalized attention weight The calculation formula is: , where is the attention weight of collaborative vehicle i to the data source j for data collection, m is the total number of data sources, is the exponential result of the scoring function, ensuring that a higher score has a greater impact on the weight.

[0041] After calculating the attention weight, the feature data of the collaborative vehicle is weighted according to these weights to obtain the contribution value of the feature data of each collaborative update vehicle. The calculation formula for the contribution value of the feature data of each collaborative update vehicle is: , where is the contribution value of the feature data of collaborative update vehicle i, is the attention weight of collaborative vehicle i to the data source j, represents the feature data obtained by vehicle i from data source j, and m represents the total number of data sources.

[0042] Using the attention mechanism, according to the collaborative weight of each collaborative update vehicle and the weight of each data source for data collection, calculate the contribution value of the feature data of each collaborative update vehicle. Specifically, obtain the collaborative weight of each collaborative update vehicle and the weight of each data source for data collection. The collaborative weight represents the contribution degree of the collaborative vehicle in the map update process, while the weight of the data source reflects the reliability of the data provided by the data source; then, define the weight calculation method in the attention mechanism, and use the collaborative weight of the collaborative vehicle and the weight of the data source for data collection to calculate the attention weight of each data source. This weight combines the weight of the collaborative vehicle and the weight of the data source and is normalized to ensure that the contributions of each collaborative vehicle and data source in the map update are allocated proportionally; subsequently, calculate the contribution value of the feature data of each collaborative update vehicle. This process multiplies the attention weight by the feature data value provided by the collaborative vehicle to obtain the contribution of each vehicle's data to the final map update; finally, the contribution values of the feature data of all collaborative update vehicles will be weighted and aggregated according to the weights to ensure that the vehicle data with higher weights has a greater impact on the map update process.

[0043] In summary, the embodiments of the present application have at least the following technical effects: First, based on the current planned route of the target vehicle combined with the current driving speed, analyze the map update timing sequence of the current driving route. The map update timing sequence is the map update time window for each section of the planned route. Then, according to the current planned route and its map update timing sequence, set the collaborative connection conditions. The collaborative connection conditions include the collaborative vehicle route and the collaborative pre-time zone, and the collaborative pre-time zone is used to represent the pre-driving space-time area of the collaborative vehicle. Next, based on the collaborative connection conditions, connect the collaborative update vehicles, align the spatial positions and arrange the time sequences according to the collaborative pre-time zones of the collaborative update vehicles, and configure the collaborative weights of the collaborative update vehicles. The collaborative weights correspond to the offset differences in spatial position alignment and time sequence arrangement. Finally, according to the collaborative weights, perform feature aggregation on the obtained dynamic maps of the collaborative update vehicles according to the spatial position alignment and time sequence arrangement relationship to obtain map update data, and use the map update data for dynamic map update. This solves the technical problems of poor real-time performance and inaccurate data in high-precision map updates of existing driverless vehicles, and achieves the technical effects of improving map update accuracy and real-time performance and enhancing the driving safety and reliability of driverless vehicles.

[0044] Embodiment 2, based on the same inventive concept as the high-precision map dynamic update method for driverless vehicles in the foregoing embodiment, as Figure 2 shown, the present application provides a high-precision map dynamic update system for driverless vehicles, wherein the system includes: Analysis module 11: Based on the current planned route of the target vehicle combined with the current driving speed, analyze the map update timing sequence of the current driving route. The map update timing sequence is the map update time window for each section of the planned route; Condition setting module 12: According to the current planned route and its map update timing sequence, set the collaborative connection conditions. The collaborative connection conditions include the collaborative vehicle route and the collaborative pre-time zone, and the collaborative pre-time zone is used to represent the pre-driving space-time area of the collaborative vehicle; Configuration module 13: Based on the collaborative connection conditions, connect the collaborative update vehicles, align the spatial positions and arrange the time sequences according to the collaborative pre-time zones of the collaborative update vehicles, and configure the collaborative weights of the collaborative update vehicles. The collaborative weights correspond to the offset differences in spatial position alignment and time sequence arrangement; Map update module 14: According to the collaborative weights, perform feature aggregation on the obtained dynamic maps of the collaborative update vehicles according to the spatial position alignment and time sequence arrangement relationship to obtain map update data, and use the map update data for dynamic map update.

[0045] Further, the analysis module 11 is used to execute the following method: Extract the section nodes of the current planned route and identify the length of each section; calculate the estimated passing time of each section node according to the current driving speed and the length of each section; perform chronological arrangement of the section nodes according to the estimated passing time of each section node in chronological order and the time interval distance to obtain the map update time sequence of the current driving route.

[0046] Further, the analysis module 11 is used to execute the following method: Calculate the adjusted time threshold of each section node according to the length of the section and the current driving speed; set the window tolerance duration according to the change penalty coefficient of each section node in the current planned route, and the change penalty coefficient is determined according to the section length and the route adjustment difficulty; perform constraint adjustment on the adjusted time threshold according to the window tolerance duration to determine the map update time window of each section node and obtain the map update time sequence of the current driving route.

[0047] Further, the configuration module 13 is used to execute the following method: Obtain the pre-space position and pre-time relationship between the collaborative update vehicle and the target vehicle; align the coordinates of the target vehicle and the collaborative update vehicle according to the pre-space position to determine the position alignment amount; calculate the time deviation amount between the collaborative update vehicle and the target vehicle according to the pre-time relationship; calculate the spatial alignment weight and the time alignment weight according to the position alignment amount and the time deviation amount, and perform weight fusion according to the preset fusion coefficient to obtain the collaborative weight value of the collaborative update vehicle.

[0048] Further, the configuration module 13 is used to execute the following method: Obtain the collaborative weight value of the collaborative update vehicle, and its calculation expression is: , where is the collaborative weight value of the collaborative update vehicle i, is the fusion coefficient of the spatial alignment weight, is the fusion coefficient of the time alignment weight, is the position alignment deviation amount, is the time deviation amount, is the spatial alignment attenuation factor, is the time alignment attenuation factor.

[0049] Further, the configuration module 13 is used to execute the following method: Obtain the record update database of the collaborative update vehicle. The record update database is the historical collection data of the collaborative update vehicle, including different collection device tags; perform data credibility evaluation according to the record update database to obtain the data evaluation coefficients of each collection device; configure the collaborative weights of the collection data sources of the collaborative update vehicle according to the data evaluation coefficients of each collection device, and add the collaborative weights of the collection data sources to the collaborative weights of the collaborative update vehicle.

[0050] Further, the map update module 14 is used to execute the following method: Obtain the dynamic map feature data of the collaborative update vehicle. The dynamic map feature data has a data source identifier; align the spatial positions and arrange the time series of the dynamic map feature data, and use the collaborative weights obtained by spatial alignment and time alignment to perform feature screening on the dynamic map feature data to obtain the map spatio-temporal features to be aggregated; calculate the feature data contribution values of each collaborative update vehicle through an attention mechanism according to the collaborative weights of the collection data sources and the collaborative weights of the collaborative update vehicle; perform feature aggregation on the map spatio-temporal features to be aggregated according to the feature data contribution values of each collaborative update vehicle to obtain the map update data.

[0051] It should be noted that the above sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above describes specific embodiments of this specification. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0052] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included within the protection scope of the present application.

[0053] This specification and the drawings are only exemplary descriptions of the present application and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A method for dynamically updating a high-precision map of a driverless vehicle, characterized in that, The method includes: Analyzing the map update time sequence of the current driving route according to the current planned route of the target vehicle in combination with the current driving speed, where the map update time sequence is the map update time window for each section of the planned route; Setting collaborative connection conditions according to the current planned route and its map update time sequence, where the collaborative connection conditions include collaborative vehicle routes and collaborative pre-time zones, and the collaborative pre-time zones are used to represent the pre-driving space-time regions of collaborative vehicles; Connecting collaborative update vehicles based on the collaborative connection conditions, aligning the spatial positions and arranging the time sequences according to the collaborative pre-time zones of the collaborative update vehicles, and configuring the collaborative weights of the collaborative update vehicles, where the collaborative weights correspond to the offset differences in spatial position alignment and time sequence arrangement; According to the collaborative weights, aggregating the features of the dynamic maps obtained by the collaborative update vehicles according to the spatial position alignment and time sequence arrangement relationship to obtain map update data, and using the map update data for dynamic map update.

2. The high-precision map dynamic update method for an autonomous vehicle according to claim 1, wherein The analyzing the map update time sequence of the current driving route according to the current planned route of the target vehicle in combination with the current driving speed includes: Extracting the section nodes of the current planned route and identifying the length of each section; Calculating the estimated passing time of each section node according to the current driving speed and the length of each section; Arranging the section nodes in time sequence according to the estimated passing time of each section node and the time interval distance to obtain the map update time sequence of the current driving route.

3. The high-precision map dynamic update method for an autonomous vehicle according to claim 2, wherein, The obtaining the map update time sequence of the current driving route includes: Calculating the adjusted time threshold of each section node according to the length of the section and the current driving speed; Setting the window tolerance duration according to the change penalty coefficient of each section node in the current planned route, where the change penalty coefficient is determined according to the section length and the route adjustment difficulty; Constraining and adjusting the adjusted time threshold according to the window tolerance duration to determine the map update time window of each section node and obtain the map update time sequence of the current driving route.

4. The high-precision map dynamic update method for an autonomous vehicle according to claim 1, wherein, The connecting collaborative update vehicles based on the collaborative connection conditions, aligning the spatial positions and arranging the time sequences according to the collaborative pre-time zones of the collaborative update vehicles, and configuring the collaborative weights of the collaborative update vehicles includes: Obtaining the pre-spatial position and pre-time relationship between the collaborative update vehicle and the target vehicle; Aligning the coordinates of the target vehicle and the collaborative update vehicle according to the pre-spatial position to determine the position alignment amount; Calculating the time deviation amount between the collaborative update vehicle and the target vehicle according to the pre-time relationship; Calculating the spatial alignment weight and the time alignment weight according to the position alignment amount and the time deviation amount, and performing weight fusion according to a preset fusion coefficient to obtain the collaborative weight of the collaborative update vehicle.

5. The high-precision map dynamic update method for an autonomous vehicle according to claim 4, wherein, The obtaining the collaborative weight of the collaborative update vehicle has the following calculation expression: , where is the collaborative weight for co-updating vehicle i, is the fusion coefficient of the spatial alignment weight, is the fusion coefficient of the temporal alignment weight, is the position alignment deviation, is the time deviation, is the spatial alignment attenuation factor, is the temporal alignment attenuation factor; Among them, the spatial alignment attenuation factor is used to adjust the influence of the position alignment deviation amount, and determines the attenuation speed of the position deviation amount on the collaborative weight. The larger the attenuation factor, the more sensitive the influence of the position alignment deviation amount on the weight, the greater the position deviation, and the faster the weight decreases; the time alignment attenuation factor is used to adjust the influence of the time deviation amount, and determines the attenuation speed of the time deviation amount on the collaborative weight. The larger the attenuation factor, the more sensitive the influence of the time deviation amount on the weight, the greater the time deviation, and the faster the weight decreases.

6. The high-precision map dynamic update method for an autonomous vehicle according to claim 4, wherein, Obtaining the collaborative weight of the collaborative update vehicle further includes: Obtaining the record update database of the collaborative update vehicle, where the record update database is the historical acquisition data of the collaborative update vehicle, including different acquisition device tags; Conducting data credibility evaluation based on the record update database to obtain the data evaluation coefficients of each acquisition device; Configuring the collaborative weight of the acquisition data source of the collaborative update vehicle according to the data evaluation coefficients of each acquisition device, and adding the collaborative weight of the acquisition data source to the collaborative weight of the collaborative update vehicle.

7. The high-precision map dynamic update method for an autonomous vehicle according to claim 6, characterized in that, According to the collaborative weight, performing feature aggregation on the obtained dynamic map of the collaborative update vehicle according to the spatial position alignment and the time sequence arrangement relationship, to obtain map update data, including: Obtaining the dynamic map feature data of the collaborative update vehicle, where the dynamic map feature data has a data source identifier; Performing spatial position alignment and time sequence arrangement on the dynamic map feature data, and using the collaborative weight obtained by spatial alignment and time alignment to perform feature screening on the dynamic map feature data to obtain the map spatio-temporal features to be aggregated; Calculating the feature data contribution value of each collaborative update vehicle through an attention mechanism according to the collaborative weight of the acquisition data source and the collaborative weight of the collaborative update vehicle, where the feature data of the vehicle refers to the relevant information provided by each collaborative update vehicle during map dynamic update; the attention mechanism automatically adjusts the influence of the feature data of each vehicle according to the collaborative weight of each collaborative update vehicle and the collaborative weight of the acquisition data source; Performing feature aggregation on the map spatio-temporal features to be aggregated according to the feature data contribution value of each collaborative update vehicle to obtain the map update data.

8. High-precision map dynamic update system for driverless vehicles, characterized in that, A system for implementing the high-precision map dynamic update method for an autonomous vehicle according to any one of claims 1-7, the system includes: Analysis module: Analyzing the map update time sequence of the current driving route according to the current planned route of the target vehicle in combination with the current driving speed, where the map update time sequence is the map update time window for each section of the planned route; Condition setting module: Setting collaborative connection conditions according to the current planned route and its map update time sequence, where the collaborative connection conditions include collaborative vehicle routes and collaborative pre-zone, and the collaborative pre-zone is used to represent the pre-driving spatio-temporal area of the collaborative vehicle; Configuration module: Connecting collaborative update vehicles based on the collaborative connection conditions, performing spatial position alignment and time sequence arrangement according to the collaborative pre-zone of the collaborative update vehicle, and configuring the collaborative weight of the collaborative update vehicle, where the collaborative weight corresponds to the offset difference of spatial position alignment and time sequence arrangement; Map update module: According to the collaborative weight, perform feature aggregation on the obtained dynamic map of the collaborative update vehicle according to the spatial position alignment and the temporal sequence relationship to obtain map update data, and use the map update data to perform dynamic map update.

Images